Using Machine Learning Models to Predict Asthma Hospitalizations Based on Air Pollutants

This study uses a Decision Tree regression model to analyze the relationship between air pollution (PM2.5, NO2, O3) and asthma hospitalizations in New York City, revealing significant prediction challenges and highlighting the potential for targeted healthcare interventions based on regional health vulnerabilities.

This paper explores the use of a Decision Tree regression model to predict asthma hospitalizations across 42 regions of New York City based on air pollution data. The study focuses on three key air pollutants: PM2.5, NO2, and O3. The model’s performance is evaluated using Mean Absolute Error, Mean Absolute Percentage Error, R-Squared Score, and Root Mean Squared Error. While the study does not aim to produce definitive forecasts, it assesses the viability of this machine learning approach for identifying health vulnerabilities related to air pollution. Forecasting the asthma hospitalizations by region can help identify and rectify inequities in health vulnerability. The model achieved a Mean Absolute Error of 50.01 and a Mean Absolute Percentage Error of 84.3%, indicating significant prediction challenges. Results show that high levels of NO2 and O3 are associated with increased asthma-related ER visits. These findings highlight potential areas for targeted healthcare interventions.

Explore More!

Research Paper